Calorie restriction: Part I – an introduction


What does restricting calories have to do with longevity?

In 1987, John Rowe and Robert Kahn, writing in Science (Rowe and Kahn, 1987), acknowledged that “a revolutionary increase in lifespan,” had already occurred. The next step in gerontology, they concluded, was increased “health span,” reportedly using this term for the first time in written history, “the maintenance of full function as nearly as possible to the end of life.”

Almost 500 years before Rowe and Kahn were calling for the quality-of-life elixir Alvise “Luigi” Cornaro believed he had bottled the brew.

As a young and wealthy Venetian merchant, Cornaro liked to party. It apparently took a toll. In his thirties, suffering from a “train of infirmities,” as he put it, Luigi sought medical attention. The prescription from his physicians? A “sober and regular life.” Basically, they told him to go on a diet. Eat less. Perhaps not surprisingly, he had already been given this advice from the authorities. Equally unsurprising, Cornaro “could not put up with it.” But after getting back on the horse—and staying on it—over the course of a year, he found himself “freed from all … complaints.”

Limiting himself to 350 grams of food (consisting of soup with an egg yolk, bread, and meat, fish, or fowl) and 414 mL1 of wine a day, he became “exceeding healthy.” Cornaro went on to publish Discourses on the Sober Life (clearly not a book on teetotaling) to share his wisdom.

He noted that people would live long “if, as they advanced in years, they lessened the quantity of their food.” Cornaro was practicing what some of us now call caloric restriction (CR) in the literature. (Note that when we refer to CR, it’s “caloric restriction without malnutrition,” unless otherwise stated.) CR is typically defined relative to the person’s previous intake before intentionally restricting calories, or relative to an average person of similar body type, or relative to the person’s estimated total daily energy expenditure (TDEE). For example, if it’s estimated that a male expends 2,000 kcals per day on average, he is practicing 30% CR if he eats 1400 kcals per day.

“Others say,” writes Cornaro in 1558, “that temperance may indeed keep a man in health, but that it cannot prolong his life.” In other words, perhaps CR can increase healthspan, but surely it can’t increase lifespan? “To this, I answer, that experience proves the contrary; and that I myself am a living instance of it.” (Cornaro reportedly died at age 98.)

This whole CR thing was nothing new, Cornaro noted (in the mid 16th century, mind you). Galen, the Greek physician—born in 129 AD—was extolling the same virtues. The idea has stood the test of time. Ben Franklin wrote in Poor Richard’s Almanac in 1733, “To lengthen thy life, lessen thy meals.” Thomas Edison wrote of Cornaro’s thesis in a blurb for Moody’s Magazine in 1916: “I have for fifty years carried out the idea of Luigi Cornaro. My forefathers had the same characteristic, and lived beyond one hundred.” (Edison died at 84.)

It was only in the last century that CR was put to the test in more experimental settings. Clive McCay more formally kicked things off in the early 1930s (McCay et al., 1935), describing that CR without malnutrition prolongs mean (i.e., how long a population lives on average) and maximal lifespan2 in rats compared to ad libitum (i.e., given free access to food) feeding.

Today, CR is argued as the most robust intervention into biologic aging in experimental animals. Perhaps we’re being a little selfish, but humans are the species (or experimental animals, as it were) of interest. What we want to know is whether CR can extend lifespan and healthspan in us. Researchers are working on the answer. The Comprehensive Assessment of Long-term Effects of Reducing Intake of Energy (CALERIE) study is one such investigation. The “L” for “Long-term” was two years of sustained CR in humans, so they weren’t exactly looking at hard endpoints (e.g., mortality).

Should we care that investigators mimicked CR in a single-celled yeast and showed an extension in lifespan (Jiang et al., 2000Lin et al., 2000)? In isolation, probably not. Taken in context, it may be a useful piece of knowledge that can help us determine whether we’re onto something bigger (from our selfish perspective). If we see similar effects in a diverse range of species that span a remarkable amount of time (about a billion years of evolution from the time yeast arrived on the scene to mammals), it increases the chances that what we’re looking at in a petri dish, or cage, is relevant to humans. In other words, if an intervention can increase lifespan and/or healthspan across all eukaryotic kingdoms, it’s probably more relevant to humans than an intervention which only works in one or two species. (The compound rapamycin, for example, is the only known pharmacological agent to extend lifespan from yeast to mammals. This will be discussed in far greater detail in a future post.)

One of the biggest pieces of evidence for proponents of CR extending life in people is that yeast (Jiang et al., 2000Lin et al., 2000), flies (Chippindale et al., 1993), worms (Klass, 1977), rotifers (Fanestil and Barrows, 1965), fish (Comfort, 1963), spiders (Austad, 1989), beetles (Ernsting and Isaaks, 1991), hamsters (Stuchlikova et al., 1975), water striders (Kaitala, 1991), protozoans (Rudzinska, 1951), water fleas (Lynch and Ennis, 1983), grasshoppers (Hatle et al., 2006), rats (McCay et al., 1935), mice (Weindruch et al., 1986), dogs (Kealy et al., 2002), and monkeys (Bodkin et al., 2003Colman et al., 2009) can reportedly live longer when subjected to CR. It should be noted that not all species respond to CR by living longer, according to the literature. Some species, and even strains within the same species can respond differently to similar CR protocols. The DBA/2 mouse’s life is typically shortened (Forster et al., 2003) by CR and Mediterranean fruit flies don’t live longer on CR (Carey et al., 2002), for example. In a study of 41 different strains of mice, investigators found that dietary restriction shortened lifespan in more strains than it lengthened lifespan (Liao et al., 2009). Steven Austad and his colleagues found that wild-derived mice showed little to no response to CR in terms of lifespan (Harper et al., 2006). There’s also a lot more to the story of the research on CR in monkeys (Mattison et al., 2012) than simply concluding CR extends life in primates, but this story is so important and interesting that it deserves an entire post of its own. More on that in the coming months.

At 30,000 feet, it appeared foolproof that CR extends life. However, as Richard Feynman famously said of science at Caltech’s 1974 commencement address, “The first principle is that you must not fool yourself-and you are the easiest person to fool.” A major thrust of Nerd Safari is to question just about everything, including ourselves and our biases. We’re trying to add the feature that is typically missing in Cargo Cult Science, to borrow the title of Feynman’s address. Taking a rigorous look at the issues, imagining all of the different ways in which the researchers (and us) might be wrong or only partially right. Looking for counterexamples and alternative explanations for the things that we’re seeing or not seeing in the literature and experiments. Trying to spot things that would make an observation or experiment invalid.

For whatever reason, what’s often missing or minimized in scientific publications are the limitations of the study, which is arguably the most important part of science. Learning from our mistakes is one of the most powerful ways to improve ourselves and build knowledge. Many observational studies, for example, read as arguments for the prosecution, when in science it is one’s own cross-examination that is perhaps the most valuable information and most deserving of ink on one of the pages of the approximately 28,000 peer-reviewed journals out there.

With that in mind, what other factors besides “CR” could explain the results? Roger McDonald, a professor of nutrition at UC Davis, and Jon Ramsey, a professor and researcher of molecular biosciences at the same institution, write in 2010, “There was no question now that McCay was correct in his original suggestion in 1935 that energy per se was the only nutritional factor causing life extension observed during long-term CR.” However, five years later, Luigi Fontana, the co-director of the Longevity Research Program at Washington University in St. Louis, and Linda Partridge, a professor and director at University College London’s Institute of Healthy Ageing,3consciously avoided using the term CR in a collaboration in Cell (Fontana and Partridge, 2015), describing the promotion of health and longevity through diet. They noted, until recently, reduced intake of overall calories was considered primary to any health benefit conferred from such “CR” studies. The authors believe that this assumption was based on a flawed interpretation of experimental data in rats, and argued that “subsequent studies in yeast, invertebrate model organisms and rodents has instead clearly demonstrated that a reduction in specific nutrients in the diet, rather than reduced calorie intake, is primarily responsible for improvements in health and extended lifespan, which is why we use the term DR [dietary restriction] rather than CR [calorie restriction].”

A 2017 study in Cell Metabolism (Roberts et al., 2017) entitled, “A Ketogenic Diet Extends Longevity and Healthspan in Adult Mice,” certainly lends to the idea that restriction in particular macronutrients (i.e., DR) may “mimic” the effects of CR… or is it the other way around? Interestingly, Jon Ramsey, who we met in a previous paragraph closing the book on nutritional factors outside of CR causing life extension, perhaps realized it might be time to reopen it: he was a corresponding author of this 2017 experiment extending lifespan with a ketogenic diet in mice. “The results surprised me a little,” Ramsey told UC Davis. “We expected some differences, but I was impressed by the magnitude we observed—a 13 percent increase in median life span for the mice on a high-fat versus high-carb diet. In humans, that would be seven to 10 years. But equally important, those mice retained quality of health in later life.”

“My thinking has changed since 2010,” Ramsey told us (via email). “I had thought decreasing energy intake was the sole driver of life span extension with CR.” Ramsey continued:

Despite the studies by Morris Ross showing that life span in CR animals could be modulated by changing macronutrient balance [Ross, 1959Ross, 1961], I predicted that life span would not be different between diet groups in a well controlled CR study using a long-lived mouse strain. My dietary fat source and life span study in CR mice [Lopez-Dominguez et al., 2015] made me change my thinking that energy intake was the only factor influencing life span with CR. The ketogenic diet study [Roberts et al., 2017] also showed that diet macronutrient composition has a significant impact on life span in mice maintained on slight CR (all diet groups were maintained at the same energy intake). I think there is sufficient information to suggest that fatty acids, amino acids and ketones could impact pathways that influence aging in ways distinct from a simple decrease in energy intake. Nonetheless, I still think energy has the greatest impact on life span.

There have been studies suggesting that protein (Ross 1959Ross 1961Solon-Biet et al., 2014) or methionine restriction extends lifespan in rats (Orentreich et al., 1993) and mice (Miller et al., 2005). Other studies, like the one Ramsey conducted (Roberts et al., 2017Davis et al., 1983) with his colleagues, suggest that higher protein intake did not decrease lifespan, and the investigators noted that no increase in lifespan was found in rats fed diets when protein was reduced to levels similar to their own study (Nakagawa and Masana, 1971Ross and Bras, 1973). Is it protein, carbohydrate, fat, total kcals, ketones, or something else that’s extending or shortening lifespan?

What else can we learn from the animal data? Does it make a difference that experimental studies are conducted in the laboratory with animals in captivity that would otherwise live quite differently outside? What can mice tell us of men? What do DR studies in non-human primates (e.g., rhesus monkeys) demonstrate in terms of diet, disease, and longevity? What about the quality of experimental diets and how closely they resemble what the particular animals eat in the wild? Does meal timing, meal frequency, macronutrient composition, circadian clocks, microbiome, nutrient sensing, sleep deprivation, age of onset of the intervention, genetic manipulation and variation, developmental programming, and other variables, play any role? Does the optimal macronutrient composition differ between species, and between animals within the same species that eat an unrestricted diet compared to one that is not? (Can someone please hand me Occam’s razor?)

What about human data? Is there anything we can learn about CR or DR from existing populations? If we’re trying to determine if DR increases healthspan and lifespan, what can the more than 300,000 centenarians4 living on the planet tell us? More than any other factor, this 0.4% of Americans may tell us that picking your parents is the best way to increase your chances of living past 100 years of age. However, environmental factors—and how those factors interact with our genes—plays a big role in delaying chronic diseases associated with aging. Modifying your genes is not currently on the menu, but diet certainly is, and it may be possible to find clues from certain regions relatively devoid of chronic diseases that live healthfully to old ages. In addition, can any of the human trials tell us anything meaningful, like the aforementioned CALERIE study, which was designed to determine the biological effects of two years of CR? Can we extrapolate anything from studies like the Minnesota Starvation Experiment or the Biosphere 2?

Can we explain the mechanisms by which CR extends lifespan? When we reduce calories, we reduce specific macronutrients. When we reduce specific macronutrients, do we have to necessarily reduce calories (although if we reduce one, we have to compensate by increasing the amount of another macronutrient, which certainly adds to the complexity of the research)? We have nutrient sensing mechanisms sensitive to amino acids, glucose, and fatty acids, but do we have calorie sensors, per se? Digging into the collective insights from the cellular and molecular landscape of longevity, and pathways involved in DR, may provide us with a better understanding of the mechanisms at play, and where aging intersects at the molecular level.

“Research,” said Nobel Prize winner Albert Szent-Györgyi, “is to see what everybody has seen and think what nobody has thought.” Because we can’t do prospective, randomized experiments with humans over a long enough period of time to definitively know if different forms of dietary restriction increase human longevity, we have to settle for the next best thing. We argue that the next best thing is a triangulation of major insights—none of them individually strong enough to be the complete basis of our understanding, but collectively much stronger, by focusing on their overlapping implications—which not only give us an insight into what to do today, but, equally important, offer a model to consider tomorrow’s treatments. “It is therefore of first-rate importance that you know how to triangulate,” said Feynman. “That is, to know how to figure something out from what you already know. It is absolutely necessary.”


Stay tuned for future installments (click here for Part IIA – monkey studies) where we’ll dig deeper into DR and attempt to answer as many questions as we can. Or as Feynman might suggest, let’s see if we can figure something out from what we already know.


* This article is a repost which originally appeared on PETER ATTIA MD.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.